Whenever cheating allegations stagger the global chess community, the world’s leading “chess detective” answers the call to offer his investigative nous.
And he does it between university lectures donning a Buffalo Bills cap.
Kenneth Regan, an associate professor in UB’s Department of Computer Science, has spent the past month wading through the “paranoia” gripping the chess world after a series of cheating allegations and renewed scrutiny over integrity in the sport.
On Sept. 5, Magnus Carlsen, the reigning world champion, raised eyebrows after pulling out of the tournament following a loss and resigned from a game in another tournament on Monday after just one move.
His opponent in both instances, and the current subject of swirling cheating allegations, is 19-year-old American Grandmaster (GM) Hans Niemann.
An International Master (IM) of chess himself, Regan is the premier authority for the International Chess Federation (FIDE) on all things tournament integrity. Regan, who analyzed more than 200 of Niemann’s games since 2020, says he found “no evidence of cheating.”
On Saturday, Regan sat down for an interview with The Spectrum to offer a glimpse at his investigative stakeout and dissect the latest cheating crisis threatening to upheave the chess world.
Drawing a bead
After losing against Niemann and withdrawing from the Sinquefield Cup, Carlsen tweeted a clip of soccer manager José Mourinho saying: “I prefer not to speak; if I speak I am in big trouble.”
The move, unprecedented in Carlsen’s career, immediately sparked furor and speculation around the then-unspoken implications of cheating targeting the young American. The allegations have since become explicit, with Carlsen clarifying his cryptic tweet in a statement.
“I know that my actions have frustrated many in the chess community,” Carlsen said. “I’m frustrated… I believe that cheating in chess is a big deal and an existential threat to the game.
“I believe that Niemann has cheated more — and more recently — than he has publicly admitted.”
Following Carlsen’s departure from the Sinquefield Cup, Niemann admitted to two counts of cheating in online chess, once aged 12 and once more aged 16 in a post-match interview with GM Alejandro Ramirez. However, the American grandmaster maintains that as the extent of his infraction.
Chess.com, the online chess platform on which Niemann played, announced in a Sept. 8 statement that it had removed Niemann’s account in light of “information that contradicts his statements regarding the amount and seriousness of his cheating.”
Niemann maintains that he has never cheated over the board or in a “tournament with prize money.”
The row between Carlsen and Niemann represents the most prominent cheating scandal rocking the chess world since the 2006 World Chess Championship, when GM Topalov’s camp accused his challenger, GM Kramnik of receiving computer assistance while taking an “unreasonable number of trips” to the bathroom during the match.
Dubbed “toilet-gate,” it was the first incident that put Regan’s name on international notice and paved the way for the emergence of his now-FIDE-recognized cheat detection methodology.
Once touted as a successor to legendary American grandmaster Bobby Fischer, Regan made his start at the chessboard at age 10. Regan was rated at around 1400 ELO by the age of 11. He shot to over 2100, just weeks before his 13th birthday. Regan achieved his master title the year after, one of the youngest since Fischer to do so.
The professor is now back in the dugout, with similar speculation of computer-assisted cheating churning around Niemann.
Commentators and online spectators alike have spawned a myriad of explanations for the American’s victory over Carlsen, including the use of vibrating anal beads to wirelessly communicate the best moves — a theory fanned by a since-deleted Elon Musk tweet.
Spectators also suggested that Carlsen’s opening preparation was leaked. Meanwhile, others have suggested Niemann simply may have simply performed better that day.
Regan reiterates that he has not yet found any evidence of cheating using his model, which he first began developing following the 2006 “toilet-gate” debacle.
The professor uses a “garden-variety predictive analytic model” with two stages. Regan first compares results between players with similar ELO ratings and computes the deviation from the mean.
“Flipping a coin 100 times, 50 is the expectation, and the standard deviation is five,” Regan said.
From there, Regan runs “a predictive analytic model” which generates a probability on every possible outcome of an event or decision.
It’s a model Regan says is applicable to every walk of life, from modeling consumer behavior when choosing a toothbrush brand to insurance companies modeling the probabilities of home damage caused by natural disasters.
“In my case, the probabilities are of a person of a certain rating, choosing [top engine] chess moves in a given position,” Regan said. “I do this in a way that not only gives me probabilities but also gives me confidence intervals, so it allows me to project the number of agreements over a range of positions and a series of games.” The model then provides Regan with a 95% confidence interval and a z-score to judge the probability of deviations beyond the normal range.
The conclusion, however, comes together with a dash of human reasoning.
“There are a lot of standout moves for Niemann to make,” Regan said. “The game was easy to play. And, yes, he matches that test 71%, which is about 10 points higher than someone in Carlsen’s class would usually match, but the projection was 69%.”
The result was well within the parameters that Regan and FIDE established to determine fair play.
“Niemann played well — one standard deviation up — but by definition, the standard deviation happens,” Regan said.
Regan acknowledged that his model would not be able to account for speculation of Carlsen’s leaked preparation for his Sinquefield encounter with Niemann. And he says that Carlsen’s 10th move was a novelty — a move that had never been played before in the game’s current position, according to existing chess databases.
“Niemann, however, predicted that move and his own crucial reply three or four moves, before the game,” Regan said. “In his post-match press conference, Niemann called that a ridiculous miracle that he happened to guess how the game would go, including four moves after the novelty.
“It’s possible that Carlsen found that rather too much of a miracle.”
Regan also addressed rumors that Niemann might’ve cheated using a computer, including one that alleges Niemann used a wireless anal-bead communicator.
Regan said that was not outside of the realm of possibility.
“One person had a buzzing device on their thighs — and you may have caught a reference that people thought the buzzing device could be in a region… that’s not as exposed on the surface of the skin,” Regan said, bemused.“There has been use of earpieces: There has been a case of a person with a hidden camera in a necktie that was giving the accomplice a view of the board.”
However, he urges the public to refer back to statistical analysis and science when in doubt.
“[Speculation] is not substantial, and it’s very much the kinds of innuendo without scientific documentation that happens all too often,” Regan said.
The professor also says that patience is required when different scientific models yield differing results.
Pawnalyze.com, a chess analysis website, conducted its own probability testing based on the points won by Niemann in tournament play rather than evaluating his moves against the computer.
The site found that Niemann’s rating rose more rapidly than anyone else of his caliber. He significantly outperformed his ELO rating of 2484, as of January 2021.
“I get completely different results from my own measure,” Regan said.
After running 106 data points of Neimann’s over-the-board and online games over the time period through his own methodology, Regan found that Niemann’s performances still essentially manifested as a normal distribution.
Laurels of patience
The science and models trying to understand whether or not Niemann cheated have broader implications for social behavior, according to Regan.
“This sets up an interesting scientific divergence,” Regan said. “I’m measuring directly the agreement with the computer and the amount of error on moves that don’t agree — something related directly to the cheating mechanism. Whereas Pawnalyze is measuring something on the output — how many points did he score? — And reaches an opposite conclusion.
“This is a societal case, where you could reach completely different conclusions based on the instrument of measurement that you choose.”
The professor stresses the importance of patience and willingness to slowly work through the calculations before drawing a conclusion.
“You have to look under the hood for an explanation,” Regan explained. “And it’s going to take a lot more work actually going over all the tournaments, not just this one, to smooth that out. But that takes a lot of work.In our thirty-second sound bite and five-second-click information age, nobody has the patience to do the work.”
He cites public outcry over the COVID-19 pandemic and doomsday fears of a recession as examples of tendencies to err toward conspiracy — and the chess world is no exception.
“There are many, many cases like this one,” he said. “Maybe in chess, I’ll have the opportunity to get to the bottom of one of them.”
Kyle Nguyen is a senior news/features editor and can be reached at email@example.com
Kyle Nguyen is a senior news/features editor at The Spectrum.